The Global Index on Responsible AI (GIRAI) is a comparative study that evaluates countries on their national policies, human rights protections, and institutional capacity related to artificial intelligence. Produced by the Global Center on AI Governance, a research organization based in South Africa, the index aims to provide a standardized measurement of how governments worldwide approach responsible AI development and deployment. GIRAI draws on extensive data collection and expert review to generate country-level scores that inform policy discussions and international comparisons.
First Edition (2024)
The inaugural edition of GIRAI was released in June 2024 and covered 138 countries and jurisdictions. The Netherlands achieved the highest score with 86.16 out of 100. This score was derived from more than 1,800 questions per country answered by local researchers, generating over two million data points collected between November 1, 2021, and November 1, 2023. Before publication, these data points were reviewed by outside experts from various organizations to ensure accuracy and consistency. Other top-performing countries in this edition included Germany, Ireland, the United Kingdom, and the United States.
Second Edition (2026)
A second edition of GIRAI was published in 2026, expanding its assessment to 135 countries and jurisdictions. This edition utilized more than 68,000 data points covering legislation, policy, government implementation, civil-society activity, and underlying conditions such as rule of law, labour rights, cybersecurity, and access to information. Norway received the highest overall score with 75.26 out of 100, followed by Italy (72.71), Ireland (71.39), France (70.32), and the Netherlands (69.51). The global average score was approximately 35 out of 100, indicating significant variation in responsible AI governance across nations.
Regional Highlights in 2026
In the 2026 edition, Nigeria ranked thirty-eighth globally but scored the highest in Africa with 45.93 points, followed by Egypt. In East and Southeast Asia, Japan achieved the highest score, followed by South Korea, Singapore, China, and Hong Kong. These regional results highlight disparities in AI governance capacity and policy development across different parts of the world, with some emerging economies showing stronger performance relative to their regional peers.
Policy Frameworks on Misinformation and Violence
The 2026 edition also identified regional differences in the adoption of policy frameworks addressing AI-driven misinformation and violence. This finding underscores the growing concern about AI's societal impacts and the varying readiness of governments to regulate these issues. The index's attention to such topics reflects its broader focus on human rights and responsible AI practices, beyond purely technical or economic considerations.
Methodology and Purpose
GIRAI's methodology relies on structured questionnaires completed by local researchers, followed by external expert review to validate findings. The index covers both formal policy documents and practical implementation, as well as civil-society engagement and enabling conditions like legal infrastructure. By providing comparable data across countries, GIRAI serves as a tool for researchers, policymakers, and advocates to assess progress and identify gaps in responsible AI governance. Its production by the Global Center on AI Governance positions it within a growing ecosystem of AI assessment frameworks, alongside efforts from organizations such as OpenAI, Anthropic, and Google DeepMind in the private sector, and academic institutions like MIT CSAIL and Stanford AI Lab.